arXiv cs.CL· Evan Li, Krishna Deshpande, Wen Cheng, Shyamnath Gollakota·· 10 小时前AI 评分58
Mobile AI Factory：为手机定制专用小语言模型的智能体系统
Mobile AI Factory: Building Specialized Language Models for Your Mobile Device
AI 导读
Mobile AI Factory 是一个智能体系统，可根据任务说明和目标手机自动构建经过微调与量化的端侧小语言模型，由编排 LLM 决定调超参、挖掘真实数据还是生成合成样本，并用固定规则回滚损害精度的改动。
正文 · 原文
Abstract:Small language models (SLMs) running on mobile devices can keep user data private and work without connectivity. But they perform well only when fine-tuned for the target task and sized for the target device, a process that today requires repeated expert tuning. We present {\it Mobile AI Factory}, an agentic system that builds a fine-tuned, quantized SLM for a given task and phone. Given a task specification and a target phone, an orchestrator LLM chooses whether to tune hyperparameters, mine real data, or generate synthetic examples, while fixed rules roll back changes that hurt accuracy. The factory searches from small to large models across base models and quantizations, adapts its accuracy goal as it searches, and uses real measurements of the target phone across inference engines and backends. We evaluate Mobile AI Factory on ten tasks and four Android phones. On an unseen testset, its on-device models outperform two five-shot frontier models on five tasks, by up to 0.34, and match them on three more. Its search comes within 0.03 of exhaustively training all 27 model variants on an entity extraction task, while using 19$\times$ less GPU time. Phone measurements also matter: the best configuration changes across phones. Without phone measurements, Claude Opus 5.5, given the full hardware and model specifications, misses the app's requirements in 25\% of easy and moderate cases and in 60\% of hard ones. Code is available at: this https URL.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2611.13509 [cs.CL] |
| (or arXiv:2611.13509v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2611.13509 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Wen Cheng [view email]
[v1]
Tue, 10 Nov 2026 03:29:29 UTC (283 KB)
来源：arXiv cs.CL · arxiv.org